Further, open databases will potentially make it harder for iterative change on that data.
Structures and versioning become harder when that is distributed at the protocol/standard level. Right now the web is already distributed, the protocols/standards are open and versioning data/content/apps/endpoints on top of that is easy, but updating protocols/standards takes time. Apps are a second layer to that, versioning is harder as you have to roll out updates and the OSs/standards they are built on are the slower changing part, web3 is even harder as updates and versioning of the protocol/standards AND the data/content on top will need to propagate and there may be pushback and splits/forks as we are seeing ETH being difficult to change core protocols/flows. There may be innovations on this as we go but also turbulence until that is realized.
For some areas like personal data and ledger data not changing much is good. For anything beyond that it makes it harder to change. Versioning and iterative change is already difficult in some cases when a company has full control over the structures. Getting multiple companies to agree on standards is harder, especially when that impacts revenues on those systems. The same will happen with web3 maybe more intensely. Right now even open standards are actively killed off because they share data. Right now "web2" could have shared databases, and there are some, but companies push away from that and actively look to own data which is a problem. However the same will happen with web3 at additional levels including the protocols/standards now.
What may happen is APIs public facades/interfaces/signatures are more atomic/stable and less changing which is always nice to have, with the guts of the structures being more keyed/document data formats. I am always a fan of iterative change that doesn't nuke the public interfaces/facades unless absolutely necessary. It may lead to "cleaner" more generic public interfaces potentially, but it may also lead to stagnate iterations and infighting like current web standards. In a way, web3 is more about standards than what is built on it. Standards can be flexible or not, both with their pros and cons.
In a way with cryptocurrency, the real goal of the investors of web3 is recreating the web with a sort of toll system, that can be good and bad but their aim is collecting on actions on new protocols/standards.
Taking control away from companies for important data like personal data and shared public data may be good, but there is also some trouble ahead and many things to work out. The same type of sharks that capture that data are looking to go lower in the stack and control that area, and extracting fees/tolls on the movement in that structure even if it is an open database. There are pros and cons to all of this.